SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Agricola
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Related Collections
Right arrow Global Change
Right arrow Upscaling
Right arrow Nitrogen
Published in Soil Sci. Soc. Am. J. 68:1285-1294 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-4—SOIL FERTILITY & PLANT NUTRITION

Nitrous Oxide Emissions from Agricultural Toposequences in Alberta and Saskatchewan

R. C. Izaurraldea,*, R. L. Lemkeb, T. W. Goddardc, B. McConkeyb and Z. Zhangc

a Joint Global Change Research Institute, Univ. of Maryland, 8400 Baltimore Ave., Suite 201, College Park, MD 20740-2496
b Agriculture and Agri-Food Canada, Semiarid Prairie Agricultural Research Centre, P.O. Box 1030, Airport Road, Swift Current, SK S9H 3X2, Canada
c Alberta Agriculture, Food and Rural Development, 206 J.G. O'Donoghue Bldg., 7000 113 St., Edmonton, AB T6H 5T6, Canada

* Corresponding author (cesar.izaurralde{at}pnl.gov).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 NOTES
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Nitrous oxide fluxes from soils are inherently variable in time and space. An improved understanding of this variability is needed to make accurate estimates of N2O fluxes at a regional scale. The objectives of this work were to (i) characterize the influence of soil–landscape combinations and N application rates on N2O emissions and to (ii) determine the contribution of these influences on the estimation of N2O emissions at the field scale. We used static chambers and gas chromatography methods to measure N2O fluxes and collected ancillary data (mineral N, water soluble C, soil water content, soil temperature) in Canada at Mundare (AB) in the aspen parkland ecoregion and at Swift Current (SK) in the short-grass prairie ecoregion. At Mundare, measurements were taken in 1995 and 1996 by landscape position and land use. At Swift Current, data were collected in 1999 and 2000 by landscape position and N rate. At Mundare, landscape position affected N2O emissions but the pattern varied seasonally. During a 46-d period in summer 1995, a flux of 430 g N2O-N ha–1 measured in a backslope was greater than the 60 g N2O-N ha–1 measured on average in shoulder and depressional areas. The flux pattern changed during a 43-d spring thaw of 1996 when fluxes from depressional areas were greatest (1710 g N2O-N ha–1). Nitrous oxide emissions from natural areas were small. The emission pattern during summer 1996 was similar to that of 1995 but the fluxes were an order of magnitude larger. At Swift Current, N2O fluxes in summer 1999 were affected by topography and N rate. Fluxes were greatest in depressional areas receiving N at 110 kg ha–1 (3140 g N2O-N ha–1). Use of the area fraction occupied by each landscape position to calculate N2O flux increased the estimates of N2O fluxes at the field scale in five out of six cases. Further research of N2O fluxes in variable landscapes should help elucidate factors controlling N2O fluxes from pedon to field scale and thus translate into improved flux estimates at regional scales.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
NOTES
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE CONCENTRATION OF N2O in the atmosphere, estimated at 2.68 x 10–2 mL L–1 around 1750, has increased by about 17% as a result of human alterations of the global N cycle (IPCC, 2001). Nitrous oxide is a potent greenhouse gas with much greater global warming potential than CO2. When N2O reaches the stratosphere, most of it is converted to N2 through a photolytic reaction that converts O3 into O2 thereby causing the stratosphere to lose some of its shielding properties against ultra violet rays (Schlesinger, 1997). Nitrous oxide forms in soils primarily during the process of denitrification (Robertson and Tiedje, 1987) and, to a lesser extent, during nitrification (Tortoso and Hutchinson, 1990). Global annual N2O emissions from agricultural soils have been estimated to range between 1.9 and 4.2 Tg N, with about half arising from anthropogenic sources (IPCC, 2001).

The methodology proposed by Mosier et al. (1998) relates N2O emissions to the agricultural N cycle through the use of readily available FAO databases. In an effort to test the use of this methodology at a regional scale, Lemke et al. (1998a) compared seasonal fluxes of measured N2O at six sites in Alberta (Canada) against estimates made using the Mosier et al. (1998) procedure. Although the estimated fluxes compared fairly well with the measured data, the procedure does not consider site-specific characteristics (e.g., soil texture) that may strongly influence N2O fluxes.

As a ratifying country of the Kyoto Protocol,1 Canada is actively pursuing the development of protocols to estimate greenhouse gas emissions from agriculture at a national scale, including N2O emissions from soils. Brierley and Patterson (2002) recently advanced a hierarchical ecological framework being developed in Canada to scale up soil emissions of greenhouse gases from field to national levels. The framework classifies all lands hierarchically—from small- to large-scale spatial resolution—into ecozones, ecoregions, ecodistricts, and soil landscape polygons. Thus, predictions (through models or measurements) of N2O fluxes at the soil landscape level will be useful for scaling these fluxes to regional and national level.

Nitrous oxide fluxes from soils are conspicuously variable in time and space. An improved understanding of this variability is needed to improve estimates of N2O fluxes at field and regional scales. The variability in N2O fluxes can be partly explained by the complex interactions that exist among key edaphic factors (e.g., available N, water status, soil texture, and temperature) and management controls (e.g., amount and type of crop residues added to soil, and fertilizer N management) (Aulakh et al., 1992). Temporally, emissions of N2O from boreal agroecosystems in the Canadian Prairie concentrate mainly during two periods of the year: spring snowmelt and early summer (Nyborg et al., 1997; Lemke et al., 1998b). Since the spring snowmelt event can account for up to 70% of the annual N2O flux (Lemke et al., 1998a), it is highly recommended that this period be included in the calculation of annual fluxes of N2O emissions.

Several researchers have recognized the importance that landscape position (topography) has on N2O emissions (Pennock et al., 1992; van Kessel et al., 1993, Corré et al., 1996). In general, these investigators found that emissions of N2O were greater in footslopes than in shoulder positions and concluded that topography—through the regulation of the hydrological cycle—was the field level control regulating N2O emissions at the microscale level. None of this research, however, made a specific quantification of the proportional contributions of N2O fluxes from the various landscape elements to the field level. This quantification would aid in the development of robust procedures to scale up emissions of N2O from field to regional scales. Therefore, the objectives of this study were to (i) characterize the influence of soil–landscape combinations and N application rates on N2O emissions and to (ii) determine the contribution of these influences on the estimation of N2O emissions at the field scale.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 NOTES
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Description of Study Sites
Two sites were selected to conduct gas-sampling campaigns and collect ancillary data. The first site was located near Mundare, AB, within the Aspen Parkland Ecoregion in a field of the Parkland Agriculture Research Initiative demonstration farm. The soils of the region are Black Chernozems (Haplustolls) (72%) and Black Solonetz (Natrustolls) (19%) (Soil Classification Working Group, 1998; Soil Survey Staff, 1998). The soils are mostly clay loam and loam in texture and lie on low relief hummocky topography (Macmillan and Pettapiece, 2000). Salient climate and soil characteristics of the Mundare site are presented in Table 1. The field was part of a precision farming project during 1995 and 1996 (Goddard et al., 1996) and digital elevation data gathered during the study were used to select the gas sampling sites. The field is managed with commercial equipment and was direct seeded (no till) to field peas (Pisum sativum L. cv. ‘Highlight’) in 1995 and to spring wheat (Triticum aestivum L. cv. ‘Invader’) fertilized with 60 kg N ha–1 and 30 kg P2O5 ha–1 in 1996. Before 1995 the field had been cropped to canola (Brassica napus L.) in 1993 and to spring wheat in 1994. The sampling campaigns at Mundare were conducted during summer 1995, spring snowmelt 1996 and summer 1996.


View this table:
[in this window]
[in a new window]
  		Table 1. Climate, soil, and landscape characteristics of the two study sites.

 
The second site was located near Swift Current, SK, located in the Short-Grass Prairie Ecoregion. Soils in this region are mostly moderately calcareous and medium textured, lying on undulating or hummocky topography. The upper slope positions are dominated by Regosols (Entisols), while the mid and lower slope positions are dominated by Brown Chernozems (Haplustolls), although Solonetzic (Natrustolls) associations are also common. For the past several decades, this site had been maintained primarily under a wheat–fallow rotation with modest fertilizer N applications (0–40 kg N ha–1) and no addition of manure. The site was managed with commercial equipment and was cropped to spring wheat (cv. ‘A.C. Barrie’) during 1999. Gas sampling campaigns were conducted during the frost-free period of 1999 and the spring-thaw period of 2000. Treatments selected for gas sampling consisted of three rates (0, 45, and 110 kg N ha–1) of N fertilizer.

Sampling Design
We followed the landscape segmentation approach proposed by Pennock et al. (1987) to select sampling sites across landscape positions (e.g., upper, mid, lower). Landform segmentation was performed on a digital elevation model derived from differentially corrected (local base station) GPS data. This segmentation procedure was used by Pennock et al. (1992), van Kessel et al. (1993), and Corré et al. (1996) to study spatial controls on denitrification. An overview and example of this approach can also be found in Pennock and Corré (2001).

At Mundare, a representative hummock-depression combination with full soil profile characterization (Goddard et al., 1996) was selected and the entire landscape divided into four segments: shoulder, backslope, footslope, and depression. Three sampling units (replications) within each landscape position were selected for gas and ancillary measurements. Four other sites were selected for gas sampling that represented common land cover conditions in the study area: cattail (Typha latifolia) in a depressional area, alfalfa (Medicago sativa L.), native grassland, and forest (Populus spp.), the last three on relatively flat backslopes.

At Swift Current, the typical landscape selected contained only three segments: shoulder, backslope, and footslope. The N treatments were applied in strips that spanned the landscape elements (i.e., each strip included shoulder, backslope, and footslope elements). The strips were randomized within a constrained area of the field to keep variability within the landscape elements to a minimum. Three sampling units (replications) were selected at each landscape position at Mundare; four were selected at Swift Current.

Gas Sampling
The gas sampling campaigns and ancillary data collection followed broadly those described by Lemke et al. (1998a)(1998b) and Lemke et al. (1999). Briefly, the sampling procedure used vented soil chambers (12 cm in diameter and 13.5 cm in height) similar to those described by Hutchinson and Mosier (1981). At Mundare, the chambers were pushed into the soil to approximately a 2- to 3-cm depth between crop rows at each sampling time. Gas samples were drawn from the headspace at the beginning of the collection period and after 1 h by fully filling disposable 30-mL polypropylene syringes and injecting the full volume into pre-evacuated 22-mL vacutainers. At Swift Current, fixed collars were inserted 3 cm into the soil between crop rows. A vented chamber (15 cm in diameter and 10 cm in height) was then clamped to this collar during the collection period. A butyl rubber O-ring was used to seal the interface between collar and chamber. Samples were drawn from the headspace after 30 min by fully filling disposable 20-mL polypropylene syringes and injecting the full volume into pre-evacuated 13-mL exetainers. Time zero values were estimated using a method similar to that described by Anthony et al. (1995). A series of ambient air samples were collected at each sampling time. The mean of these samples was used as the time zero concentration. The change in concentration was calculated by subtracting the time zero concentration from the final concentration.

Nitrous oxide concentration does not necessarily change linearly with time during deployment of static vented chambers (Hutchinson and Mosier 1981; Anthony et al., 1995). Thus, the fluxes reported in this study may have been underestimated. Comparisons made using data from several similar sites showed that flux estimates calculated from a single time-point were generally between 80 and 100% of the estimates calculated from multiple time-points (R.L. Lemke, unpublished data, 2002).

Gas samples were collected between 1200 and 1500 h. At Mundare, samples were collected at least twice monthly during the growing season, with more frequent sampling when N2O emissions were most probable, after rainfall (high soil-water contents) and fertilizer N application. Samples were collected at least twice weekly during the spring thaw period. At Swift Current, samples were collected at least twice weekly through early spring and summer, and then with diminishing frequency through late summer and early fall, when soils were dry and emissions negligible.

The concentration of N2O in the air samples was determined with a gas chromatograph (Varian Canada Inc. Mississauga, ON) equipped with an electron capture detector (ECD). Operating conditions for samples from the Mundare site were: 0.3 m poropak Q precolumn, 1.5 m poropak QS analytical column; carrier gas Ar/CH4 (95:5 ratio) at 30 mL min–1; injector set at 100°C, columns at 45°C, and ECD at 400°C. Operating conditions for samples from the Swift Current site were: 0.45 m poropak N precolumn, 1.83 m Haysep D analytical column; carrier gas Ar/CH4 (90:10 ratio) at 30 mL min–1; injector set at 100°C, columns at 80°C, and ECD at 380°C. Calibration curves for N2O were prepared by serial dilution from a 5.0 µL L–1 custom standard (Praxair Specialty gases, Edmonton, AB).

Seasonal and annual cumulative N2O-N loss estimates were calculated for each plot (replicate). Cumulative loss estimates were estimated by linear interpolation between data points and integrating the area under the flux versus time curve.

Ancillary Measurements
At Mundare, two soil cores (0–10 cm) were collected from each gas sampling location on 13 Mar. 1996 (before spring thaw), 25 Apr. 1996 (after the spring thaw event), and 30 May 1996 (beginning of the growing season). These were the only sampling dates for this study site. The two cores were combined and then crushed and mixed thoroughly by hand; extractions were performed on field-moist basis within 24 h. At Swift Current, soil samples (0- to 15-, 30- to 60-, 60- to 90-, and 90- to 120-cm depths) were taken in 1999 on 26 May, 17 June, 8 July, 29 July, 25 August, and 30 September to determine available N, water soil organic C and soil moisture. The soil samples were passed through a 2-mm sieve and stored air-dry until processing. Samples were also collected monthly during the growing season from 0 to 7.5 and 7.5 to 15 cm soil depths to determine available NO3–N and NH4–N, water-soluble organic carbon (WSOC) and soil moisture. Additional samples (0- to 10-cm depth) were collected at approximately weekly intervals during May and June to further monitor soil NO3–N and NH4–N status. These samples were handled the same as the Mundare samples described above. For all samples, soil NO–3–N and NH+4–N were extracted by shaking the sample in a 2 M KCl solution (extractant/soil = 10:1) for 1 h, and then filtering (Whatman No. 42, Whatman Ltd., Maidstone, UK). The filtrate was analyzed with a Technicon Autoanalyzer II (Technicon Industrial Systems, Tarrytown, NY). Water soluble organic C was extracted by shaking the sample (extractant/soil = 10:1) in distilled water for 1 h on a reciprocating shaker set at 140 excursions per minute, letting the solution settle for 15 min then centrifuging for 30 min at 12500 x g. The supernatant was filtered through a 0.45-µm Millipore (Millipore Corp., Bedford, MA), type HA filter using a standard Millipore filter clamp apparatus with vacuum, directly into a 100-mL poly sample container. The filtrates were stored at –10°C until analysis was performed. The Swift Current samples were analyzed for total dissolved organic carbon (DOC) using a Dohrmann DC-180 Carbon Analyzer (Tekmar-Dohrmann, Mason, OH). The Mundare samples were analyzed for DOC using an Astro 2001 Series 2 Soluble Carbon Analyzer (Astro Inter. Corp., League City, TX). The water content of the soil samples was established by gravimetric determination of weight loss when soil samples were dried at 105°C for 24 h. Water-filled pore space (m3 m–3) was calculated as the ratio between volumetric soil water content (m3 m–3) and total porosity (m3 m–3). Volumetric soil water content was obtained by multiplying gravimetric soil water content by soil bulk density ({rho}b, Mg m–3) values reported in Table 2. Total porosity was calculated as (1 – {rho}b/{rho}s) where {rho}s is soil particle density assumed to be 2.65 Mg m–3.


View this table:
[in this window]
[in a new window]
  		Table 2. Soil characteristics by landscape position and soil horizon.

 
Statistical Analyses
Analyses of variance were conducted on N2O-N flux data using ANOVA and GLM procedures in SAS version 8.2 (SAS Institute Inc., Cary, NC). REG and when required by normality tests (UNIVARIATE procedure), the N2O-N flux data were log-transformed. After mean separation, the log-transformed means of cumulative fluxes were back transformed to their original scale for presentation (Davis, 1986). Ancillary data are reported in terms of their means and standard errors of the mean (MEANS procedure). Multiple regression analyses of N2O fluxes vs. available N (NH4–N + NO3–N), WSOC, N rate, gravimetric soil water content, and water-filled pore space were conducted with the backward elimination option in PROC REG. The level of probability for a variable to stay in the regression was set at 10%.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 NOTES
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
The results of the N2O field measurements reported and discussed here were obtained at two sites of hummocky topography, which are typical of glacial landscapes (Macmillan and Pettapiece, 2000). As indicated in Table 1, however, the climate at Mundare is more humid and cooler than that at Swift Current. The soils at Mundare are genetically more developed than those at Swift Current, as suggested by the lack of a B horizon at the latter site (Table 2). The Ap horizons at Mundare contain roughly double the amount of organic C than do those at Swift Current.

Nitrous Oxide Fluxes at Mundare
An average air temperature of 14.7°C during the first gas-measurement period (11 July 1995–25 Aug. 1995) was 1.0°C cooler than the long-term average measured during the July–August period at Vegreville (AB), located about 16 km east of the study site. A total precipitation of 190 mm recorded during the same period was 31% greater than that of the normal period. An average air temperature of –0.9°C during 15 Mar. 1996–26 Apr. 1996, the second period of gas measurements, was 1.0°C warmer than the normal. A period precipitation of 43 mm was near normal. An average air temperature of 14.3°C during the third gas-measurement period (29 May 1996–24 Sept. 1996) was 0.4°C warmer than the June–September normal. The site received 25% more precipitation during this period than that received during a normal year.

Landscape position affected cumulative N2O emissions but the pattern varied seasonally (Table 3). During a 46-d period in summer 1995, the backslope position emitted five to seven times more N2O than the other three positions. Due to variability in the measurements, however, the N2O flux from the backslope position was statistically higher than those from shoulder and depressional areas but not from footslope positions (Table 3). The pattern of N2O emissions by landscape position changed during a 43-d spring thaw in 1996 when N2O fluxes from depressional areas were significantly greater than the other three positions. In general, the emission pattern observed during the 119-d measurement period of summer 1996 was similar to that of 1995 but the fluxes were an order of magnitude larger (Table 3). There were no statistical differences in N2O emissions among the three lower positions (backslope, footslope, and depression). The N2O flux from the shoulder position was similar to that of the depression position in summer 1995 and to the backslope position during spring thaw 1996 (Table 3). For comparison during this period, N2O emissions from natural (forest, grass) or perennial cover (alfalfa) areas were small during the 46-d measurement period of summer 1995 (approximately 40 g N2O-N ha–1) except for typha-covered depressions where the N2O-N flux (500 g ha–1) was similar to that of backslope positions in the agricultural field nearby (400 g ha–1). What are the implications to these observations? While we cannot separate with certainty the contribution of vegetation and slope position to the N2O flux, this "vegetation cover" comparison seems to be consistent with the observations made on the cultivated site. In the absence of N additions, moisture regime (mediated by slope position) appears to be the dominant controlling factor.


View this table:
[in this window]
[in a new window]
  		Table 3. Cumulative N2O-N flux (g ha–1) at Mundare (AB) at different landscape positions during 1995–1996.

 
In addition to different climatic conditions, cropping patterns may have also influenced the seasonal N2O emission patterns observed during the summers of 1995 and 1996 (Table 3). As stated above in the Materials and Methods section, the Mundare site was cropped to field peas in 1995 and to spring wheat in 1996. Recent work by Lemke et al. (2002) and Wagner-Riddle and Thurtell (1998) suggest that N2O emissions could be low during the pulse phase. We surmise that the 60 kg ha–1 of fertilizer N added as urea to the wheat crop, in combination with the N released from the residue decomposition of the pulse crop, likely stimulated N2O emissions (Aulakh et al., 1992).

As shown in Table 3, N2O emissions during the spring thaw of 1996 were larger than those observed in the summer of 1995. Nitrous oxide emissions were detected soon after the beginning of the thaw period (mid March 1996) (Fig. 1). However, no N2O fluxes were detected during the cold spell that occurred during late March and early April when air temperatures declined abruptly to around –20°C. Emissions of N2O resumed soon after average air temperature increased to about 5°C. Nitrate-N concentrations in footslope and depression positions in samples taken before the thaw (13 Mar. 1996) were three to five times higher than those in shoulder and backslope positions (Table 4). Nitrate levels in the two lower slope positions were substantially lower in the post-thaw soil samples taken on 25 Apr. 1996. Nyborg et al. (1997) observed a similar effect on soil samples taken in late fall and early spring. Both available N and soil water content at the beginning of the thaw period appear to be good predictors of N2O flux (Table 4). Individually, initial soil NO–3 could explain about 46% of the variability in cumulative N2O loss. However, multiple regression analysis with the backward elimination procedure in SAS revealed gravimetric soil water content (GSWC) to be a better N2O-flux predictor than available N (N2O-N [g ha–1] = –755 + 3855 GSWC [kg kg–1], R2 = 0.729**).



View larger version (18K):
[in this window]
[in a new window]
  		Fig. 1. Nitrous oxide fluxes at Mundare, AB, from four landscape positions during the spring thaw period of 1996. Also shown are average air temperatures during the measurement periods and, with black arrows, days with significant precipitation ≥5 mm.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Mean available soil N, water soluble C, and soil water content (0- to 10-cm depth) at Mundare at various sampling times during 1996.

 
Nitrous Oxide Fluxes at Swift Current
Precipitation was near normal and well distributed during early summer (April–July) of 1999, although the late summer and fall period was dry. Cumulative N2O fluxes were influenced by both topography and N-fertilizer rate during a 128-d period in summer 1999 (26 May–30 September) but not during a 45-d period in spring thaw 2000 (3 March–26 April) (Table 5). In 1999, plots receiving 45 kg N ha–1 (recommended rate) did not emit more N2O than plots receiving no N fertilizer. However, applying N at 110 kg ha–1, a rate well in excess of recommended levels, did significantly increase N2O flux compared with the no N treatment. Cumulative N2O flux from the shoulder positions was significantly lower than from that from footslope positions. When comparisons were made within the same fertilizer N rate, the influence of landscape on N2O fluxes in summer 1999 was only statistically significant at the highest N rate where the footslope position had higher losses than the other two positions. Cumulative N2O fluxes during spring thaw 2000 were smaller than those from the previous summer. At an N rate of 110 kg ha–1, the backslope evolved larger amounts of N2O than shoulder and footslope positions.


View this table:
[in this window]
[in a new window]
  		Table 5. Cumulative N2O-N fluxes (g ha–1) at Swift Current (Saskatchewan) at different landscape positions and three N rates during 1999–2000.

 
When comparisons were made within the same landscape position in 1999, fertilizer application increased N2O emissions in footslopes (Table 5). Fluxes of N2O from backslopes receiving N at 110 kg ha–1 were greater than unfertilized backslopes and similar to those from backslopes receiving N at 45 kg ha–1. Fluxes of N2O from shoulders receiving N at 110 kg ha–1 were greater than those from the other two rates (0 and 45 kg N ha–1). Fluxes of N2O during spring thaw 2000 were low and unresponsive to either slope position or N rate except for the backslope position fertilized with 110 kg N ha–1 (Table 5). Our data revealed that applying fertilizer N in excess of crop requirements leads to significant increases of N2O emissions, especially from footslope positions.

Available N (NH4–N + NO3–N) remained low (<6 mg N kg–1) during the growing season of 1999 and did not reflect any influence either from the N application rate or the landscape position (Table 6). Levels of WSOC remained fairly stable over the growing season and revealed little sensitivity to treatment effects. Gravimetric soil water content decreased during the growing season and was highest on footslope positions.


View this table:
[in this window]
[in a new window]
  		Table 6. Available soil N, water soluble C, and soil water content at Swift Current at various sampling times during 1999.

 
Within the same fertilizer N rate, spring wheat yields were comparable across the landscape positions (e.g., at 110 kg N ha–1, spring wheat yielded 2.62 Mg ha–1 in the footslope position while it yielded 2.56 Mg ha–1 in the shoulder position). In contrast, grain yield did increase with increasing levels of N application when compared within the same landscape element. However, when fertilizer N was applied at 110 kg ha–1 the relative increase in N2O emissions was greater than the relative increase in grain yield on both the backslope and footslope positions. This can be demonstrated by expressing cumulative N2O-N loss as a percentage of grain N harvested for each N rate-landscape element combination (Table 7). We chose to express the results in this way because N is a unit common to the inputs (fertilizer N), losses (N2O-N), and to crop yield (grain N harvested). Lemke et al. (2002) have argued that expressing N2O emissions as a percentage of harvested N provides a useful way to compare the "N2O intensity" of growing disparate crops (e.g., cereals versus pulses) or farming systems (e.g., forages versus annual grain production). When compared within landscape element, N2O expressed as a percentage of harvested N for the 110 kg N ha–1 rate was significantly higher than the 0 and 45 kg N ha–1 rate for the backslope position, and significantly higher than the 0 kg N ha–1 rate on the backslope position. When compared within the same fertilizer N rate, N2O expressed as a percentage of harvested N was significantly higher in the footslope compared with shoulder position, but only for the highest rate of N applied. Our results underscore the need to match the amount of fertilizer N applied to crop N needs, particularly for those areas of the landscape with soil–water regimes that favor N2O production.


View this table:
[in this window]
[in a new window]
  		Table 7. Comparison of N2O losses expressed as a percentage of N harvested at Swift Current in 1999.

 
Multiple regression analysis of N2O flux vs. soil variables (available N, water soluble organic C, gravimetric soil water content, and water-filled pore space) and N rate were conducted by breaking up the summer period of 1999 into five subperiods according to soil sampling times (Table 6): 26 May–21 June, 22 June–8 July, 9 July–29 July, 30 July–25 August, and 26 August–30 September. The two variables that explained most of the variation in the flux data were water-filled pore space and N rate (Table 8). These results are similar to those reported by Linn and Doran (1984) in tilled and no-tilled soils. The flux of N2O increased exponentially as water-filled pore space increased, especially after 0.5 m3 m–3 (Fig. 2). However, the variability in the fluxes also increased suggesting that other factors were at play.


View this table:
[in this window]
[in a new window]
  		Table 8. Regression coefficients explaining N2O flux for five measurement periods during the summer of 1999 at Swift Current, SK.

 


View larger version (17K):
[in this window]
[in a new window]
  		Fig. 2. Relationship between N2O flux (g NO2–N ha–1) and water-filled pore space (m3 m–3) at Swift Current during the summer of 1999.

 
The lack of a clear relationship between N2O emissions and soil mineral N has been reported by other workers (Matson and Vitousek, 1990; Davidson and Hackler, 1994; Mosier et al., 1996; Corré et al., 1999). These authors have shown that N2O production is more strongly related to N turnover than to mineral N pool size. Fertilizer N rate is a better indicator of N turnover than point-in-time measurements of soil mineral N status. Conversely, N2O production following snowmelt may show a stronger relationship with the existing mineral N pool because soil mineralization–immobilization processes (N turnover) are likely limited at this time.

Nitrous oxide emissions were low during the spring thaw of 2000. Estimated N2O loss ranged between 23 and 431 g N2O-N ha–1 during this 59 d (4 March–26 April) sampling campaign. Comparison of main effects showed that cumulative N2O losses were not significantly influenced by either hillslope position or the rate of fertilizer N applied in the previous growing season. When comparisons were made within the same fertilizer N rate, hillslope position was only significant at the 110 kg N ha–1 rate where emissions from the backslope were significantly higher than from shoulder and footslope positions. Soil NO–3 levels at the Swift Current site were very small (<5.0 mg NO3–N kg–1) on all plots in the fall of 1999 (Table 6), as they were in frozen soil cores (0–15 cm depth) collected on 3 March, just before the spring thaw event (data not shown). Multiple regression analysis of N2O flux vs. soil variables (available N, WSOC, gravimetric soil water content, and water-filled pore space) and N rate (0 and 100 kg N ha–1) detected WSOC as the only significant variable in the regression equation but its predicting ability was low [N2O-N (g ha–1) = –116 + 4.9 WSOC (kg kg–1), R2 = 0.321*].

Lemke et al. (1998a) found a relatively strong correlation between late fall surface soil-NO–3 levels and cumulative N2O loss during the following spring, and inferred from their data that soil NO–3–N levels in the range of 5.0 mg kg–1 or less would limit N2O production. Results from this study appear to support this hypothesis. The dry conditions in late summer and fall of 1999, combined with limited snow cover during the winter period, resulted in soil profiles that were quite dry and likely remained well aerated throughout the winter and spring thaw period in the shoulder and backslope positions. These conditions would not favor N2O production. Surface soils in the footslope position did reach and remain at saturation for several days (based on visual observation), but N2O emissions remained low. We conclude that N availability was the main factor limiting N2O production during this time period, although dry, well-aerated soils no doubt further decreased the potential for N2O loss.

Scaling-Up Nitrous Oxide Fluxes
What then are the scaling implications of our observations? Clearly, there were significant variations in N2O fluxes among hillslope positions (Tables 3 and 5). At Mundare, N2O fluxes were greatest on backslope positions during the summers of 1995 and 1996 (Table 9). During the spring thaw of 1996, however, the order changed and depressional areas exhibited the largest N2O fluxes (Table 9). Without any further information, the arithmetic average of the fluxes measured on the four hillslope positions would yield the best representation of N2O fluxes at the field scale. At Mundare, shoulder positions represented 23% of the entire field, backslope positions 40%, footslopes 25%, and depressional areas 12%. Using this information, we calculated the weighted average by season and found that these were greater than the arithmetic averages (Table 9). The weighted N2O flux during summer 1995 was 34% greater than the corresponding arithmetic average. A similar trend was obtained for summer 1996 (7%) while the weighted N2O flux during spring thaw 1996 was 30% lower than the calculated arithmetic average.


View this table:
[in this window]
[in a new window]
  		Table 9. Seasonal N2O fluxes at Mundare, fractional area occupied by landscape position, and fraction of N2O emitted by landscape position.

 
Using a similar analysis for the Swift Current site, we find that the ascending order in N2O flux was shoulder < backslope < footslope (Table 10). Although N rate did not affect this order, increasing N rate tended to interact with hillslope position to increase the proportion of fertilizer N emitted as N2O. At Swift Current, backslope positions accounted for 55% of the field, footslopes accounted for 35% and shoulders for the rest. Again, the weighted average calculation returned, in our view, a more realistic representation of N2O fluxes at the field scale. Weighted N2O flux averages were 7–34% greater than the arithmetic N2O flux except during the spring thaw at Mundare in 1996 when the weighted average flux was 34% smaller than that calculated using the arithmetic flux.


View this table:
[in this window]
[in a new window]
  		Table 10. Seasonal N2O fluxes at Swift Current; fractional area occupied by landscape position, and fraction of N2O emitted by landscape position.

 
Which approach would yield more accurate flux estimates at the field scale: the arithmetic or the landform weighted average? Clearly, we cannot determine this in this paper since we lack a flux measurement made at the field scale to compare against. In variable landscapes, however, we hypothesize that the weighted average approach would yield the most accurate estimate of N2O flux at the field level because through the quantification of the landform components one would be able to better capture (or account for) all the factors controlling N2O fluxes.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 NOTES
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
At Mundare, landscape position affected cumulative N2O emissions but the pattern varied seasonally. Air temperature, soil water content, and to a lesser extent available N controlled N2O fluxes during the spring thaw of 1996. At Swift Current, N2O fluxes were affected by landscape position and N rate. The amount of N2O lost, when expressed as a percentage of N harvested, increased with N rate in the backslope and footslope positions. At Swift Current, water-filled pore space and N rate explained >70% of the N2O flux variability during various periods in 1999. Use of the area fraction occupied by each landscape position to calculate N2O flux increased the estimates of N2O fluxes at the field scale by at least 7% in five out of six comparisons. Further research of N2O fluxes in variable landscapes should help elucidate the main factors controlling N2O fluxes from soil both in space and time and thus translate into improved flux estimates at regional scales.


   ACKNOWLEDGMENTS
 
Research supported by the Natural Science and Engineering Research Council (NSERC) of Canada, Agriculture and Agri-Food Canada, the Program of Energy Research and Development (PERD), Natural Resources Canada, and the USDA Consortium of Agricultural Soils Mitigation of Greenhouse Gases (CASMGS).


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 NOTES
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
1 Canada ratified the Kyoto Protocol on 17 Dec. 2002. Back

Received for publication May 6, 2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 NOTES
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
M. P. Dusenbury, R. E. Engel, P. R. Miller, R. L. Lemke, and R. Wallander
Nitrous oxide emissions from a Northern Great Plains soil as influenced by nitrogen management and cropping systems.
J. Environ. Qual., March 1, 2008; 37(2): 542 - 550.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
R. F. Grant, E. Pattey, T. W. Goddard, L. M. Kryzanowski, and H. Puurveen
Modeling the Effects of Fertilizer Application Rate on Nitrous Oxide Emissions
Soil Sci. Soc. Am. J., January 6, 2006; 70(1): 235 - 248.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Agricola
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Zhang, Z.
Related Collections
Right arrow Global Change
Right arrow Upscaling
Right arrow Nitrogen

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome